Motor Controller Wierdness, Battery Life, Throttle Issues.

I need to build a new motor controller. The one I repaired and upgraded off that old dead electric scooter has a (new) wierd issue with the throttle input voltage. It outputs more or less linearly to around 60-odd percent of full power, tracking along with the throttle input voltage up to around 3V (can't remember the exact voltage), then at a specific point just beyond that, it suddenly jumps up to full power.

It's not the actual throttle input device itself, as I've used different kinds, mostly pots of various types and a few Hall effect sensors, and any of these setups that outputs up to that point and beyond will give that sudden change. Anything below that and I can't even get full power.

As bad off as the original controller electronics were when I got them (multiple fried components), I'm just going to start from scratch, and wire one up based on one of the motor controller chips designed for this sort of thing, and wire that up to the 100A/100V MOSFETs and drivers I had upgraded the old controller with, which are all on their own heatsink.

Even with that heatsink just sitting there on the plate at the bottom of the Magna frame's rear triangle (the one the motor is bolted to underneath), those MOSFETs aren't getting hot, just fairly warm, even when I used the motor by itself to drive the bike at around 15MPH for about 5 minutes or so. The ambient temperature was around 75°F, so they'll probably get significantly hotter during summer. To prevent that, I'll be putting the whole controller in a box with a good heatsink bolted to the bike (or the outside of a cargo pod, by the rear wheel so the wheel's motion will also help move air past it, which worked pretty well on the upright bike's controller). I may also put a small PC-style thermally-controlled fan on it to ensure there is no overheating, with the sensor mounted on the heatsink near the MOSFETs.


Battery life seems good so far. I've run them about an hour today, in four trips--one to the grocery store on the way to work, then a second to get the rest of the way to work. A third was after work back to the same store for something I forgot, then back home the same way I came before work. Almost 10 miles in all, with no noticeable difference in power provided from start to finish (although I'm sure there is a difference). That's much better than I could get with the old 12v-12Ah batteries and the frictiondrive setup.

Life test will continue tomorrow/etc, until they're run down enough that there's not really useful power from them, so I can see how long they might last on longer trips (before I try taking one with them and being stuck without power).


The throttle knob I have is difficult to use in traffic, partly due to the lack of freewheels on the pedals/motor interface. Having to hold it in place all the time against the spring (even as light a tension as it is) is tiring, and because it's only really reachable by my thumb, just like my rear shifter, I have to let it go to shift. If I'm shifting gears for slower speeds, that's ok, but if I'm trying to shift into faster gears it's difficult to keep pedalling at the faster speed when the motor is now being a drag on the pedals, so it starts to take my attention away from the road and put it on the shifting/throttling which I need to be able to do automatically.

The only good solution I can do anytime soon is to build that tension-controlled throttle. I have some ideas that should be reasonably quick to implement and test, though they're not ideal and won't be a chain-tensioner in themselves, which is what I'd like to use so I can get rid of the derailer being used as a chain-tensioner.

The first idea is simply to use a lightly-sprung pivot arm with a rubber roller against the top of the chain, right in the center of it's top chainline where the difference between tense and not is greatest. That arm would have a magnet on it that sweeps near a hall-effect sensor, wired into the throttle. It's basically like the one I want to make, except it's not mechanically strong enough to actually be used as a tensioner. I have an arm setup out of some old high-end Sony VTR machines I was given as boxes of salvaged parts, but I almost certainly have to change the sensor on it which is some form of CCD optical sensor, as far as I can tell, and I have no idea what it's pinout or ratings/etc are (I didn't see a p/n on it). That arm was probably used in the VTR to do something very similar, to track tape tension and control motor speed, or something similar, to prevent damage to the tape.

Since there is the possibility that something will go wrong with that setup, I'll wire it in using a diode on the input line, so that I can also keep the current throttle, also wired in with a diode.

Eventually I'll get the tensioner/throttle built, and then a cable from it to the cable-style throttle twist-grip control from the Honda Spree will be able to pull the tensioner up against the spring just as if I was pedalling, should I need an override for any reason.